EXECUTIVE SUMMARY — In Plain English

Nu Holdings, better known as Nubank, is a digital bank born in Brazil that focused on one simple idea: Latin American banking was broken — expensive, slow, and unfriendly — and technology could fix it. Instead of branches and paperwork, Nu offers a mobile app where you can do everything from opening a checking account to getting a credit card or loan in minutes. Think of it as “the digital Wells Fargo of Latin America” — except cheaper, simpler, and built entirely on software.

The company makes money largely the same way any bank does — by earning interest on loans and collecting fees — but it does so on a radically lower cost base. Its customers deposit money; Nu lends some of that money out (mostly through credit cards and personal loans) and earns the spread between what it pays on deposits and what it collects on loans. It also takes small cuts from purchases made with its cards, collects interchange fees from merchants, and sells premium services (such as Ultravioleta for affluent users, and small-business accounts).

Unlike most banks, Nu has no physical branches, runs almost entirely online, and uses artificial intelligence to handle credit decisions and customer service. That allows it to operate with an efficiency ratio below 20% — meaning only $0.20 of cost per $1 of revenue, versus $0.60–$0.70 at traditional banks.

In essence, Nubank’s business model sits at the intersection of technology and finance: it attracts tens of millions of customers with free, intuitive financial tools, then steadily deepens that relationship via credit, savings, insurance, and payments. As it grows, each new product makes the customer more valuable; each new customer gives Nu more data to improve its models — a compounding dynamic similar to that seen in Amazon or MercadoLibre.

As of 2025, Nu has 131 million customers across Brazil, Mexico, and Colombia, generating $11.5B in revenue and nearly $2B in net profit. The model is simple: low-cost customer acquisition through word of mouth and digital onboarding, high engagement through app-based services, and steadily expanding monetization per customer. Growth here doesn’t just make Nu bigger — it makes it better, with scale economies, data learning loops, and regulatory strength reinforcing its moat.

1. HOW DOES NU ACTUALLY MAKE MONEY?

Transaction Walkthrough

Imagine a young professional in São Paulo opens the Nu app to get a credit card. Approval takes minutes, powered by Nu’s AI risk model. Every time she buys groceries, Nu earns:
- Interest if she carries a balance (the main profit engine)
- Interchange fees from merchants (typically ~2–3% of purchase value)
- Late fees if she misses payments
Her paycheck then lands in her Nubank account, generating deposits Nu can use to fund loans — paying customers ~10% local deposit interest but earning ~40%+ on its credit card receivables. The difference (“net interest margin”) is profit, after provisions for defaults.

On top of this, Nu earns service fees on payments (Pix, NuPay), subscription income (Ultravioleta premium card), and SME banking fees (business accounts, payment tools).

Revenue Breakdown (2024 Actuals — best available data)

Segment Details

1. Credit (65% of total)
- What it does: Issued credit cards and personal loans across Brazil, Mexico, and Colombia.
- Pricing: Interest rates per local regulations; annual percentage generally 40–60%.
- Customers: Emerging middle class, often with limited access to traditional credit.
- Competitive position: #1 in Brazil by number of credit cards; expanding fast in Mexico.
- Trajectory: High growth, but cyclical risk from consumer lending; AI underwriting stabilizes delinquency rates (90+ NPLs at ~6.6%).

2. Deposits & Float (22%)
- What it does: Paid deposits and savings accounts, generating funding base. Nu invests those funds in fixed-income securities or uses them to fund credit.
- Pricing: Pays interest to savers; earns spread via investments and lending.
- Customers: Mostly retail consumers; small business deposits growing.
- Trajectory: Rapid deposit growth ($41.9B in 2025, +29% YoY), providing low-cost funding.

3. Payments & Fees (13%)
- What it does: Processes daily transactions (Pix instant payments, NuPay for merchants).
- Revenue source: Merchant interchange and service fees.
- Trajectory: Fast-growing network effect segment — early but expanding margins.

2. WHO ARE THE CUSTOMERS AND WHY DO THEY CHOOSE NU?

• Primary base: Latin American individuals earning $500–$2,000/month; many were unbanked or underserved.
• Why Nu: No branch visit, transparent fees, app simplicity, customer service via chat. In Brazil, incumbents like Itaú and Bradesco charged high fees; Nu is free and accessible.
• Stickiness: High — Nu’s 83% activity rate and high app engagement (ARPAC up 27% YoY) show customers genuinely rely on it. Once payroll and billing are set up, switching is cumbersome.
• Lifecycle: Customers start with a card, add deposits, then loans and insurance — deepening engagement over years.

If Nu vanished, millions of users would lose primary financial access; unbanked customers have few alternatives offering similar simplicity and no hidden fees.

3. COMPETITIVE MOAT IN SIMPLE TERMS

Why competitors can’t easily copy Nu:
- Scale & data advantage: 131M customer datasets allow superior AI credit models → lower default → better pricing.
- Low-cost structure: No branches; software handles service. Traditional banks can’t match sub-20% efficiency ratio without gutting legacy systems.
- Brand trust: Nu is the most loved consumer finance app in Brazil. Switching is behaviorally and operationally costly.
- Regulatory foothold: Full banking licenses in Brazil, pending in Mexico; not trivial for new entrants.

Even a billionaire competitor would face regulatory hurdles, years of data accumulation, and brand loyalty obstacles.

4. SCALE ECONOMICS — DOES GROWTH MAKE IT BETTER OR JUST BIGGER?

Increasing Returns Confirmed:
From 2019–2024: revenue grew ~18x, net income grew from negative to $1.97B; operating costs rose far slower than revenue. Efficiency ratio fell from ~70% (est. early period) to <20%. Operating Profit CAGR > Revenue CAGR → clear increasing returns.

Mechanism: more customers = more data → lower credit losses → stronger monetization → each new dollar of revenue cheaper to earn.

Capacity Utilization Assessment

Current digital infrastructure can serve well over 200M customers (131M current). Capacity ratio ≈ 1.5–2.0x, indicating significant embedded leverage — future revenue can grow ~50–100% with little additional capex.

5. WHERE DOES THE CASH GO?

• Capital-light: CapEx only ~$175M/year — negligible versus $8.5B operating cash flow (2025).
• Reinvestment: Mostly channeled into product launches, AI, and credit portfolio expansion.
• Balance sheet: Strong — $9.9B cash against $3.5B debt.
• Capital deployment: Focused on reinvesting in growth nations (Brazil, Mexico, Colombia). No dividends yet; retains profits for scale expansion.

6. BUSINESS MODEL EVOLUTION & TRANSITIONS

Historical transition:
- 2014–2018: pure credit-card fintech.
- 2019–2021: added checking accounts, deposits, and PIX payments → became full neobank.
- 2022–2025: transitioned into multi-product financial platform, adding insurance, SME, and premium offerings.
Each step expanded monetization and stability: once cyclical lending now balanced by recurring transaction fees.

Next transition (2026–2030): global expansion + AI-as-a-banker vision. Nu’s “nuFormer” foundation model aims to personalize finances via AI — pushing it from fintech to AI-enabled financial advisor platform.

CEO David Vélez — Stanford MBA and ex-Sequoia partner — has consistently treated transitions as compounding investments, not cost cuts. Long-tenured leadership (CFO Lago, IR Souto) signals strong continuity.

7. WHAT COULD GO WRONG?

1. Credit risk: As lending scales, a macro shock (Brazilian recession) could spike defaults.
2. Regulatory tightening: Governments could cap interest rates or enforce stricter credit loss provisions.
3. AI misjudgments: Poor model calibration could misprice risk en masse.
4. Competition: Traditional banks modernizing apps or digital challengers targeting SME deposits.
5. Expansion dilution: U.S. entry could distract focus; Latin America may remain the core profit driver.

Inversion (How does it die?): A credit bubble followed by high defaults erodes trust; regulatory fine; reputation collapse among unbanked customers. Early signs would be rising NPLs and collapsing ROE — watch those closely.

BUSINESS MODEL VERDICT

In One Sentence:
Nu makes money by using its low-cost digital banking platform to lend, process payments, and earn interest on deposits — taking advantage of huge scale and data efficiencies to serve Latin America’s mass market better and cheaper than traditional banks.

Overall Assessment:
A “wonderful business” in Buffett’s sense — large, growing need, durable competitive advantages, low cost structure, and scalable returns. Nu combines fintech agility with banking economics, positioning it as one of the world’s most efficient retail financial franchises.

Bridge to Financial Analysis:
Now that we understand how Nu monetizes its platform and scales efficiently, the next step is to validate whether these structural advantages translate into sustained profitability and capital efficiency in the financials — specifically ROE, ROIC, and free cash flow trends.